PETER WADE / WORKED THINKING EXAMPLE
What does a model result actually show?
A hypothetical geochemistry claim, examined one step at a time.
The hypothetical claim: “Our PHREEQC calculation gives a positive calcite saturation index. Therefore calcite precipitated in the field and the site removed carbon dioxide.”
1. Separate the result from the conclusion
A positive calculated saturation index indicates supersaturation with respect to calcite for the specified water composition and thermodynamic data. It is a statement about the modelled solution, not a field observation of a solid forming. The USGS documentation distinguishes speciation and saturation-index calculations from simulations that explicitly include equilibrium phases.
2. Ask which assumptions control it
Check the analytical data and charge balance, the selected thermodynamic database, temperature, and whether the system exchanges carbon dioxide with a gas phase. If precipitation is simulated, check whether equilibrium was imposed and whether that is reasonable over the site’s residence time. A separate kinetic-reaction model may be needed when reaction rate matters.
3. Name the missing evidence
To argue that calcite actually formed, compare the model with field or laboratory evidence of mineral products and changes in dissolved chemistry. To argue net carbon dioxide removal, define the system boundary and carbon source, then account for gas exchange, alkalinity, dissolved inorganic carbon, solid carbonate and relevant transport or losses. A saturation index alone does not establish either outcome.
4. Make the next test discriminating
Compare open- and closed-gas assumptions, test plausible database and input uncertainty, and ask which observation would distinguish precipitation from alternative explanations. Record what each scenario predicts before deciding which measurements to collect.
The careful conclusion
“The specified solution is calculated to be supersaturated with respect to calcite. Precipitation and net carbon dioxide removal remain hypotheses requiring additional modelling and independent observations.”
Use the modelling checklist ↗